Need electronics help, PFC into Ej22t

Headlights to tailights and everything in between.

Moderators: Helpinators, Moderators

PhyrraM
Fourth Gear
Posts: 1980
Joined: Thu Aug 11, 2005 4:34 pm
Location: Lake Elsinore, CA

Need electronics help, PFC into Ej22t

Post by PhyrraM »

Mostly a personal project here, but if successful would allow a fully programmable plug and play ECU upgrade for our EJ22Ts.

As many know the main difference between EJ20G ECUs and ours is the switch from 4 channel coil on plug to 2 channel wasted spark coil packs. I'm trying to get a 4 channel ECU to drive a 2 channel ignition.

Please forgive and correct my very basic, and very new, knowledge of electronics.

From the trouble shooting section of our FSM it looks like the ecu fires the ignitor by grounding the ignition pins. The pins are held high by pull up resistors and dropped to 0 volts when the output transistors are energized. When this happens the ignitor transistors switch state and the coil discharges.

Image

The following picture is from the EJ20G 4 channel setup. There is no showing of the ECUs insides but it's safe to assume it follows the same pull-up resistor, transistor forcing it low. It appears to be the exact same set-up as the EJ22t, just with double the channels.

Image

Now here is where I need some help. I'm pretty sure you cannot just bond two ECU output pins together as this may damage one channel as the other fires? What I'm thinking might be possible is to isolate the outputs with diodes(?) before joining them together. Something like this...

Image

Thoughts? Can it be this simple? What am I missing? If it really can be done similier to this I would then research diodes to figure out what kind.
'93 Winestone SS Auto, '91 Pearl White SS.
'93 Pure White SS EJ20G slanty intercooled, SIDESWIPED! In stasis.
'94 FWD and '95 AWD Laguna Blue SVXs.
2017 Pure Red BRZ Limited w/Performance pack
vrg3
Vikash
Posts: 12517
Joined: Sun Nov 03, 2002 2:13 am
Location: USA, OH, Cleveland (sometimes visiting DC though)
Contact:

Post by vrg3 »

You're ... almost right. I'm tired right now but if I don't post an explanation tomorrow (Friday), bump this thread back up.
"Just reading vrg3's convoluted, information-packed posts made me feel better all over again." -- subyluvr2212
farfrumwork
Second Gear
Posts: 334
Joined: Thu Jan 08, 2004 1:04 am
Location: mile high

Post by farfrumwork »

That makes sense in my Mechanical Engineering head, but let's see what knowledge Vikash has to expound upon.

+eleventybillion on wanting to be able to run a PFC.
Such a simple work-around would be great!


so BUMP
Old - '93 T-Leg - 14.693 @ 92.68 at 5800ft ASL (Bandimere Speedway)
New - 07 Spec B/stg 2

"bloodstains - speed kills -fast cars - cheap thrills -rich girls - fine wine
i've lost my sense - i've lost control - i've lost my mind" Agent Orange
vrg3
Vikash
Posts: 12517
Joined: Sun Nov 03, 2002 2:13 am
Location: USA, OH, Cleveland (sometimes visiting DC though)
Contact:

Post by vrg3 »

Sigh, bad timing. This time I'm not that tired, but I am quite inebriated. Let's see how I do explaining this when drunk.

This is how a spark is created:

1) The ECU decides it is time to start charging the coil with energy, so it switches off its internal NPN transistor driving the ignitor. The pull-up resistor then provides a current source at the ignition pin of the ECU.

2) The base of the NPN transistor in the ignitor is provided with current through that pull-up, and current starts flowing through the primary circuit of the ignition coil. Energy starts to collect in the coil. This is called dwell.

3) The dwell time is up and the ECU is ready to set off the spark. It activates the internal transistor, pulling the output to ground, which cuts off current flow through the base of the transistor in the ignitor, which cuts off current flow through the primary circuit of the coil. This change causes a huge voltage on the secondary circuit, which is high enough to jump across the spark gap(s).

4) The ECU's internal transistor stays turned on until it's time to begin dwell again.

So, taking that into account, you can see that dwell is what complicates the situation and makes the diode idea not work right.

I actually did try running an EJ20G ECU on my car using four diodes as you describe once, and it barely even idled.

Sorry. Good thought, though.
"Just reading vrg3's convoluted, information-packed posts made me feel better all over again." -- subyluvr2212
PhyrraM
Fourth Gear
Posts: 1980
Joined: Thu Aug 11, 2005 4:34 pm
Location: Lake Elsinore, CA

Post by PhyrraM »

Thanks Vikash.
'93 Winestone SS Auto, '91 Pearl White SS.
'93 Pure White SS EJ20G slanty intercooled, SIDESWIPED! In stasis.
'94 FWD and '95 AWD Laguna Blue SVXs.
2017 Pure Red BRZ Limited w/Performance pack
napphappy
Second Gear
Posts: 512
Joined: Fri Jun 03, 2005 12:57 am
Location: Gig Harbor, WA

Post by napphappy »

I would recomend buying a set of usdm wrx coils. The ignitor is built in so you just have to run wires from the ecu to each coilpack.(and power and ground)

I know the bolt up perfectly to Phase II heads but you may have to modify them for ej20g heads.
1997 White BG Ltd 5MT, EJ22T with Wiescos, EJ20R Heads, Two large dents R.I.P
1995 Impreza EJ22T DOHC Hydra EMS, 20g, EWG, 3in APS TBE, AWIC
PhyrraM
Fourth Gear
Posts: 1980
Joined: Thu Aug 11, 2005 4:34 pm
Location: Lake Elsinore, CA

Post by PhyrraM »

Using Ej20G heads and ignition has been pretty well covered. We are trying to see if there is a way of using a 4 channel ECU with EJ22t heads/ignition.

There are at least 3 different plug-and-play ECUs for our cars, IF you can convert to 4 channels. Apexi Power FC, Link, and Autronic(?).

If I were to run a second ignitor circuit (which I have actually already done in prep for an EJ20G) it would be a simple matter of installing 4 small coils and 4 seperate HT leads. What I am trying to figure out is if there is a way to do it with the factory 2 channel system. This would save some expense and fabrication that may be out of some members reach.

Not looking good so far. But still have hope.
'93 Winestone SS Auto, '91 Pearl White SS.
'93 Pure White SS EJ20G slanty intercooled, SIDESWIPED! In stasis.
'94 FWD and '95 AWD Laguna Blue SVXs.
2017 Pure Red BRZ Limited w/Performance pack
Legacy777
Site Admin
Site Admin
Posts: 27884
Joined: Tue Oct 15, 2002 11:37 am
Location: Houston, Tx
Contact:

Post by Legacy777 »

Are the Apexi PFC, Link, & Autronic stand alone type ECU's, or are they piggy back chips?

Ultimately, I'd like to retain the stock ECU, and either be able to chip it and modify the info I want, or somethign similar.
Josh

surrealmirage.com/subaru
1990 Legacy (AWD, 6MT, & EJ22T Swap)
2020 Outback Limted XT

If you need to get a hold of me please email me rather then pm
PhyrraM
Fourth Gear
Posts: 1980
Joined: Thu Aug 11, 2005 4:34 pm
Location: Lake Elsinore, CA

Post by PhyrraM »

They are standalones.

The factory ECU is chipable, in fact I believe it could even be converted to a sort of programmable ECU with an EPROM emulator and the proper support software and hardware. However, as of this time it doesn't look like anybody has, or is going to crack the factory ECU. Unfortunatly this is beyond my skillset. At this point in my life I would prefer to spend the money than roll up the sleeves.

I have come up with 2 more ideas on how to use the 4 channels ECUs.

The first is adding a second ignitor circuit and swapping the 2 channel coil pack out to 4 individual coils with 4 HT leads. The Aussies do this to replace the old and often unrelaible factory coil packs. They use Bosch HEC-716 coils. I have been unable to locate these here in North America so far. I'm looking for a suitable replacement. Other than the upfront cost of the 4 coils and a second ignitor this solution has a lot of labor involved to do it right. Properly adding the 3rd and 4th channels involves cannabilizing a second harness of the proper wires then pulling the dash, HVAC and left front fender from the car to add them in to your harness. A bit more work than I was hoping to present the community with.

The second idea is to use a microcontroller to accept 4 inputs from the ECU and output 2 channels to the ignitor. It would be an Atmel AVR because I have some experiance with them and they can be programmed in BASIC (for free up to 2K). I would need some help choosing transistor types, designing protection circuits for the microcontroller and other hardware stuff like that. I believe I'm OK with construction and writing the actual program as it should be pretty straight forward. The main disadvantages for other members would be sourcing the parts, constructing the unit and getting the microcontroller programmed.
'93 Winestone SS Auto, '91 Pearl White SS.
'93 Pure White SS EJ20G slanty intercooled, SIDESWIPED! In stasis.
'94 FWD and '95 AWD Laguna Blue SVXs.
2017 Pure Red BRZ Limited w/Performance pack
vrg3
Vikash
Posts: 12517
Joined: Sun Nov 03, 2002 2:13 am
Location: USA, OH, Cleveland (sometimes visiting DC though)
Contact:

Post by vrg3 »

With the microcontroller you'd still have to determine what to do with regard to dwell. The original dwell times wouldn't be correct, so you'd have to calculate appropriate times for the coil pack. It's not necessarily that big a deal, but something to consider.

There may be another option if you are willing to install an MSD DIS-2. Since the DIS-2 operates on the principle of capacitive discharge, dwell is a non-issue. You would just have to feed it an open-collector signal with a rising edge when the spark needs to happen (which is the signal it normally gets from the ignitor).

If you were to NOR the two ignition outputs from the ECU, you'd have a falling edge whenever either channel was supposed to ignite. This signal be used as the input for an edge-triggered monostable multivibrator; the output of the one-shot would exhibit a rising edge when the spark needs to happen, and the MSD would see it and fire the spark. The one-shot would then need to reset itself before there's any chance of another cylinder needing to spark.

Off the top of my head, an NE558 seems well-suited; it's triggered by falling edges, has an open-collector normally low output, and should run happily off the car's system voltage.
"Just reading vrg3's convoluted, information-packed posts made me feel better all over again." -- subyluvr2212
Legacy777
Site Admin
Site Admin
Posts: 27884
Joined: Tue Oct 15, 2002 11:37 am
Location: Houston, Tx
Contact:

Post by Legacy777 »

Vikash......did you get my reply email to your comments about the EM upgrades?
Josh

surrealmirage.com/subaru
1990 Legacy (AWD, 6MT, & EJ22T Swap)
2020 Outback Limted XT

If you need to get a hold of me please email me rather then pm
vrg3
Vikash
Posts: 12517
Joined: Sun Nov 03, 2002 2:13 am
Location: USA, OH, Cleveland (sometimes visiting DC though)
Contact:

Post by vrg3 »

Huh, yeah, I did, and I wrote a whole reply, but apparently I forgot to hit 'send.' I'll hit 'send' now. :)
"Just reading vrg3's convoluted, information-packed posts made me feel better all over again." -- subyluvr2212
PhyrraM
Fourth Gear
Posts: 1980
Joined: Thu Aug 11, 2005 4:34 pm
Location: Lake Elsinore, CA

Post by PhyrraM »

Considering the price of a MSD DIS-2, the fact that I've already got the two extra channels in the harness and I've got an EJ20G sitting in the garage waiting I think the best way for me to proceed is with the 4 seperate coils idea.

I have found a few Bosch part numbers that pretty closely match the HEC-716 coils that the Aussies have been using to convert with. They are F3000-12127, F3000-12128 and F3000-61989. All are used in vehicles available in North America. Mostly Nissans. These should be available in most any junkyard.

I was hoping for a fast and easy fix that others could follow, but none appeared. I suppose in the larger picture, If someone is willing to pay the price of entry on just the ECU, 4 more coils and a bit of wiring is not much more.

It will be a little while before I try this, but I'll keep the thread updated when I do.
'93 Winestone SS Auto, '91 Pearl White SS.
'93 Pure White SS EJ20G slanty intercooled, SIDESWIPED! In stasis.
'94 FWD and '95 AWD Laguna Blue SVXs.
2017 Pure Red BRZ Limited w/Performance pack
SemperGuard
Second Gear
Posts: 307
Joined: Sat Sep 04, 2004 6:07 pm

Post by SemperGuard »

What if you install a 4 channel ignitor instead of a 2 channel, wire it up like stock to the ignitor, then connect the outputs together from the ignitor to the coil pack? Cyl1/2 together to front coil, and cyl3/4 together to rear coil.
PhyrraM
Fourth Gear
Posts: 1980
Joined: Thu Aug 11, 2005 4:34 pm
Location: Lake Elsinore, CA

Post by PhyrraM »

To me, it seems like that would still have the dwell problem Vikash describes.
'93 Winestone SS Auto, '91 Pearl White SS.
'93 Pure White SS EJ20G slanty intercooled, SIDESWIPED! In stasis.
'94 FWD and '95 AWD Laguna Blue SVXs.
2017 Pure Red BRZ Limited w/Performance pack
vrg3
Vikash
Posts: 12517
Joined: Sun Nov 03, 2002 2:13 am
Location: USA, OH, Cleveland (sometimes visiting DC though)
Contact:

Post by vrg3 »

Yeah, why would it matter whether you're trying to combine the signals before or after the ignitor?
"Just reading vrg3's convoluted, information-packed posts made me feel better all over again." -- subyluvr2212
SemperGuard
Second Gear
Posts: 307
Joined: Sat Sep 04, 2004 6:07 pm

Post by SemperGuard »

Well I'm not sure what minor effect dwell is having, but I can tell you first hand that using a 4 channel and connecting after the ignitor allows the engine to start, idle, and run on an EJ20G ECM with a single coil pack AKA waste spark ignition. When I say minor effect, I mean maybe instead of you getting 80% efficiency, you end up with 75%. Not noticeable to driver, but noticeable from a numbers standpoint.
vrg3
Vikash
Posts: 12517
Joined: Sun Nov 03, 2002 2:13 am
Location: USA, OH, Cleveland (sometimes visiting DC though)
Contact:

Post by vrg3 »

Huh. That's interesting.

When I combined the signals to the ignitor with diodes that was not my experience.
"Just reading vrg3's convoluted, information-packed posts made me feel better all over again." -- subyluvr2212
SemperGuard
Second Gear
Posts: 307
Joined: Sat Sep 04, 2004 6:07 pm

Post by SemperGuard »

Right, I tried the same thing at first, except no diodes. Using the stock 2 channel ignitor with #1 and 2 connected, and #3 and 4 connected before the ignitor. The engine wouldn't even start. I think the reason being is that the ECM operates differently than your thinking.

My thinking is that the transistor in the ECM is normally off, sending 12v to the transistor in the ignitor. When the ECM wants to fire a cylinder it turns on the transistor inside it, the 12v then flows through the ECM transistor, turning off the ignitor transistor and causing the coil to fire. Or something along those lines. The reason it won't work with the wires connected before the ignitor is because there is still 12v going to the ignitor from the other cylinder when it should have none...I think.

Of course with that reasoning it should still not work because the circuit isn't isolated between the ignitor and the coil, and voltage should find its way to ground without building up in the coil.

Also, which way did you wire it? The ECM I/O list isn't entirely right. #2 and #3 should be switched.
BXSS
Fourth Gear
Posts: 1301
Joined: Wed Jan 10, 2007 4:23 pm
Location: Bx, NY

Post by BXSS »

So it runs well?
If it does, can we have step by step instructions (for dummies)??

I have an Apex-i Power FC on the shelf of my garage that could go in my SS....
94 Legacy SS
96 Impreza OBS
99 Impreza OBS-T
04 Sienna XLE Limited AWD
PhyrraM
Fourth Gear
Posts: 1980
Joined: Thu Aug 11, 2005 4:34 pm
Location: Lake Elsinore, CA

Post by PhyrraM »

Alright, I finally had time to sit down and do some research on transisters and diodes. Again, forgive and correct the laymans terms.

What Semperguard is describing (4 channel ignitor directly into 2 channel coilpack) should work, AS LONG AS THE DWELLS DO NOT OVERLAP. If the factory ignitor can 'stop' the flow through the coils (diode function) during it's 'off' phase, then it should work equally as well keeping the channels seperated back at the ECU. Again, as long as the dwells do not overlap. Cylinder #1 needs to charge and fire before cylinder #2 starts to charge, even at redline.

Now off to see if I can find the dwell times for the coilpack. Most every car used a single coil, even for V8s and V12s, up until relatively recently. I'm very optimistic again.

If you've been following other threads you know I've got another SS, that is actually running. It's stock, so I need to fab a harness adapter because I don't want to do anything destructive to the factory harness. I've got an SVX ignitor on the shelf that will work just fine for these tests. Give me a week or two to cannabalize a junkyard ECU and harness to fab the adaptor. I have both a BC/BF EJ20G ecu and a EJ20G Power FC to test with.

Dead ECU and harness connector donations accepted. :smt006
'93 Winestone SS Auto, '91 Pearl White SS.
'93 Pure White SS EJ20G slanty intercooled, SIDESWIPED! In stasis.
'94 FWD and '95 AWD Laguna Blue SVXs.
2017 Pure Red BRZ Limited w/Performance pack
farfrumwork
Second Gear
Posts: 334
Joined: Thu Jan 08, 2004 1:04 am
Location: mile high

Post by farfrumwork »

nice!

I'm anxiously awaiting your findings. It does sound promising again.
Old - '93 T-Leg - 14.693 @ 92.68 at 5800ft ASL (Bandimere Speedway)
New - 07 Spec B/stg 2

"bloodstains - speed kills -fast cars - cheap thrills -rich girls - fine wine
i've lost my sense - i've lost control - i've lost my mind" Agent Orange
PhyrraM
Fourth Gear
Posts: 1980
Joined: Thu Aug 11, 2005 4:34 pm
Location: Lake Elsinore, CA

Post by PhyrraM »

It's done.

I've got a stock EJ22t, still running wasted spark and stock coil pack, running on a Legacy EJ20G ECU using a SVX ignitor. There are no mods to the car or harness. It idles and revs fine. I'm going to attempt a 30 mile freeway drive home.

Pics and details later tonight. If there are no problems, I will attempt to get it running on my Apexi EJ20G Power FC.
'93 Winestone SS Auto, '91 Pearl White SS.
'93 Pure White SS EJ20G slanty intercooled, SIDESWIPED! In stasis.
'94 FWD and '95 AWD Laguna Blue SVXs.
2017 Pure Red BRZ Limited w/Performance pack
asc_up
Fifth Gear
Posts: 2473
Joined: Sun Sep 10, 2006 9:15 am
Location: University Place, WA

Post by asc_up »

Oh my God.

This might be a huge break through!! Hahaha. I'm excited.
-Aaron

2000 Audi S4 - 2.7L Twin-turbo, 6 Speed

[quote="evolutionmovement"]It was me. And those are my balls. Happy Sunday![/quote]
PhyrraM
Fourth Gear
Posts: 1980
Joined: Thu Aug 11, 2005 4:34 pm
Location: Lake Elsinore, CA

Post by PhyrraM »

And it runs on the Power FC also. Runs well actually.

Details inbound. Dinner and shower first.

Basically, SemperGuard was right.
'93 Winestone SS Auto, '91 Pearl White SS.
'93 Pure White SS EJ20G slanty intercooled, SIDESWIPED! In stasis.
'94 FWD and '95 AWD Laguna Blue SVXs.
2017 Pure Red BRZ Limited w/Performance pack
Post Reply